Engineering leadership · Book notes
The Value Flywheel Effect from the CIO's chair
Cloud waste rose to 29% this year and regulators want exit plans, so the Wardley map in this 2022 book lands harder than its serverless-first label.
Cloud waste is rising again. The Flexera 2026 State of the Cloud survey estimates wasted cloud spend at 29% (up from 27%), marking the first increase in five years, with AI workloads cited as the primary driver. The State of FinOps 2026 report shows 78% of FinOps practices now report into the CTO or CIO organization, but most of the easy savings have already been realized.
This puts a different spin on The Value Flywheel Effect today. Author David Anderson (with Mark McCann and Michael O’Reilly) published it with IT Revolution in 2022, when people talked about cloud in terms of speed. The thesis that cost and speed follow from strategy first lands harder now that people who own the cloud bill are also being asked about the value of the technology.
The model itself is broken into four phases: clarity of purpose, challenge and landscape, next best action, and long-term value. Each phase has a different leader, with the CEO being first. Each phase is worked with a Wardley map. Hanging off of those phases are twelve tenets. Two that a platform lead might underline are “code is a liability” and “frictionless developer experience.”
This pattern comes out of the serverless-first program the authors ran at Liberty Mutual. In Tech Lead Journal, Anderson says that program reduced run costs by around 90% (his number, unaudited) and improved time to market. The map in the book does the heavy lifting: it positions components along the genesis-to-commodity axis. The rule is to build differentiators and consume commodities.
Anderson’s definition of serverless-first is more narrow than the label might imply. In the same interview, he talks about managed services as the default choice, not turning every workload into a function, and doing the simplest thing that delivers value. That makes serverless-first more of a platform decision than an architecture style.
Gartner says 80% of large software engineering organizations will have platform teams by 2026, up from 45% in 2022. The “frictionless developer experience” tenet can work as a charter for such teams: the platform makes the managed default the easy path. This is close to the cognitive-load argument in Team Topologies is an AI agent a teammate or a platform.
Banks and insurers need to push back on this part. The managed-service design is harder to move. DORA has been in effect since January 2025, requiring EU financial firms to manage ICT concentration risk and plan for exiting providers. Late last year, European supervisors named the first critical ICT third-party providers for direct oversight. So now substitutability is something the regulator can ask the CIO about.
Commodity pricing can cut the other way: 37signals says it saves $2 million/year or so after moving its compute from public cloud to on premises in 2024. But neither of those objections is addressed by the slogan. The book’s own tool can do that: a map showing which components are commodities and where lock-in is worth paying for.
The authors have advanced their argument. Now that code generation is cheap, knowing what to build trumps building fast, and clarity of purpose moves back to the front of the flywheel. But it leaves the CIO wondering: can one Wardley map be the shared artifact for the FinOps team, the platform team, and the risk function?